/robowaifu/ - DIY Robot Wives

Advancing robotics to a point where anime catgrill meidos in tiny miniskirts are a reality

Days left: 15


JulayWorld fallback document - SAVE LOCALLY

JulayWorld onion service: bhlnasxdkbaoxf4gtpbhavref7l2j3bwooes77hqcacxztkindztzrad.onion

Max message length: 32768

Drag files to upload or
click here to select them

Maximum 5 files / Maximum size: 20.00 MB

More

(used to delete files and postings)


/robowaifu/meta-2: Electric Boogaloo Robowaifu Technician 05/13/2020 (Wed) 16:22:03 No.3108
Please Note: I'm going to be continuing with occasionally repopulating our board again with old posts. Feel free to post responses to them regardless, as I will generally be watching and can often respond appropriately to your posts. But be aware these will be 'log dumps' as it were from our original board. Specifically, these bump-bot re-posts will all have rw@bump.bot as the email sig. Cheers. >--- This thread is for off-topic and general discussions. FYI, our bunker is at https://anon.cafe/robowaifu/catalog.html probably should bookmark that now Anon. :^) Also, let's discuss ways to get more people involved. How can we grow this board? As well, let's share general robowaifu ideas, etc., to help inspire each other. This thread is meant to improve /robowaifu/ in general way and be a place to hangout with loosely off-topic talk. >previous /meta thread: >>38
Edited last time by Chobitsu on 06/07/2020 (Sun) 07:18:28.
Open file (49.98 KB 550x850 the new infinity cup.png)
note: I'll go ahead and move your post here too, I-Cup Anon since you kind of got a raw deal coming in after the old thread was bump-locked. Hey, robot enthusiasts, We're revamping the Infinity Cup over at https://anon.cafe/icup/ and we would really like it if you could come and join our matches, be it as a team or even just to have fun in chat! Come see us!
>>3108 >> let's discuss ways to get more people involved. How can we grow this board? as a person who had been looking for a place like things for at least a few years now, its a relief to actually find a community that exists on furthering the accessibility and development of companion bots. That being said, as an outsider I don't think a chan IB is the best place to recruit people to considering the HEAVY stigmas going around about places like this. As much as I like the idea of being anon, I feel like participants here will immediately get dragged down alongside the false flags happening especially considering the whole mini doll fiasco that was occurring circa early to mid 2019. If anything I feel like a profile based forum like bunker would be better to gain support while avoiding any drama.
>>3115 Welcome Anon. Yes, it's a tricky situation. We currently have a recruitment thread actively going on r/n to discuss topics like this. >>2705 But truth be told, we've been at this for over 3 years now, and it's actually been a very creative experience in general--if a small-scale group thus far. To quote: >"And quite frankly, the creative benefits of us all dynamically engaging here together in a free-form way are just too important to ignore. " >>2701 >directory & glossary subsection Regardless, we're glad you finally found your way here. This would be a good place to begin to discuss your own robowaifu dreams and plans, or you can look around our catalog for yourself and pick and appropriate thread, or just start your own.
Open file (801.97 KB 2049x1019 ccp information warfare.jpg)
Open file (229.75 KB 1280x960 fracture points.png)
>>3115 I'm not against it but I wouldn't personally participate in a forum right now. It's too easy for adversaries to quantify people's weaknesses and destroy a community. It might have been okay 6 years ago but not now. They know what they're doing and don't want AI tech in the hands of common people. I'm shocked Google is even bold enough now to blatantly hide scientific research and advancements in AI. Within 1-2 years though bots will become commonplace that we cannot easily tell from a human being besides their need to control others and a lack of personal creativity and industry. We may have no other choice but to switch to a forum then unless we can filter new posters attempting to subvert the board with our own AI. It's an ongoing struggle we will have to adapt and overcome. Trying to optimize how many people we can bring in now is like hoping for fruit without involving ourselves in the process of planting and taking care of a tree. If we create something worthwhile people will come with little effort. We won't have to sell it to them. Anyway they will come and enjoy the fruit of our work with us. On the other hand trying to win people over to do the work for us is how a commie thinks. If we focus our energy too much that way we will neither have the fruit nor a tree.
>>3128 >and a lack of personal creativity and industry This will be the key Anon. Everyone needs to condition themselves to detect this now (learn to properly ask leading questions to draw that fact out if need be). >It's an ongoing struggle we will have to adapt and overcome. If you're who I think you are, then I presume you're well aware that's the one think we completely dominate at over current (or likely mid-term, say 10+ years out) AI. For us here on /robowaifu/ this is always a two-edged sword ofc. We want compelling, engaging interactions with our robowaifus, but we are more aware than most the threat that AI represents in the hands of an evil Corporate-police-state, for example. >Trying to optimize how many people we can bring in now is like hoping for fruit without involving ourselves in the process of planting and taking care of a tree. That is a great analogy Anon. Almost a beautiful one in fact. But I'm willing to let Anons explore other avenues and seek to grow our community some. We honestly do need more help after all. >If we focus our energy too much that way we will neither have the fruit nor a tree. Sober warning to us all. Thanks Anon, you really are talented with words. WE DEMAND MORE GREENTEXTS REEEEE! :^) >t. chobitsu
>>3128 >"...in case of a sudden public opinion outbreak.." ANYTHING BUT THAT! heh, fugging gommies, sheesh.
>>3128 BTW did you notice the /tism/ board being slidden right this moment? I wonder what toxic and problematic post was made that triggered it? It appears to be mostly just senseless one-word posts and strikes me as entirely automated, though the selected threads were obviously hand-picked beforehand to be obnoxious shit.
>>3131 >/tism/ check that. it wasn't spread out on the site, but instead seemed to be entirely focused on /cow/, now that i've analzyed it. could easily have been a drunk & spiteful tranny, actually. my little test didn't seem to trigger much immediate response. paranoia is a word haha
Open file (72.94 KB 1006x277 -tism- - Brave_099.png)
>>3131 >>3139 >related nearly all one-word bumps, all made within the span of just a few minutes, almost all with some sort of tranny or faggot theme in the OP.
Open file (131.08 KB 1026x492 -tism- - Brave_101.png)
>>3140 haha, i couldn't help myself. like pottery.
This may change over time ofc, but please heed the advice, Anon. https://spqrchan.xyz/.static/fallback.txt JulayWorld fallback map, v1.0 --- SAVE THIS DOCUMENT LOCALLY --- Primary domain: https://julay.world/ Backup domain in case the first one is taken down: https://spqrchan.xyz/ Onion: http://bhlnasxdkbaoxf4gtpbhavref7l2j3bwooes77hqcacxztkindztzrad.onion/ (Even if the server is lost, the onion should come back up with the same address.) If both domains are lost: https://alogs.theГунтretort.com/ (inactive until needed) In case of server loss: The images will be lost, but the text content is backed up daily. Save whatever you want to keep. Live updates during downtime: #julayworld on irc.rizon.net If everything fails, take shelter in other webring boards, that's the original purpose. If everything fucks up remember that anonymity is a spirit and not something people can just shut down. Something new will pop up, keep looking. Julay is not the first imageboard nor is it the last. --- SAVE THIS DOCUMENT LOCALLY ---
Edited last time by Chobitsu on 05/14/2020 (Thu) 06:33:16.
I figured I'd throw the topic of our CSS up for discussion/improvement here. Be aware I've learned little about CSS myself thus far, so please keep the discussion at my retarded level haha thanks. :^) >200523 current CSS: span.greentext {display: block;} span.pinktext {display: block;} .divMessage br {display: block;} body { background-image: url(https://imgur.com/K2XmKPw.png); background-repeat: no-repeat; background-position: right bottom; background-attachment: fixed; background-size: 28%; } code { white-space: pre-wrap !important; border: 1px double #e60000; background: #0d0d0d; color: #00e600; font-family: 'Lucida Console', Monaco, monospace; font-size: 105%; max-width: 100%; overflow: auto; padding: .7em 1.1em; display: block; word-wrap: break-word; } code br { display:block !important; } code .embedButton { display: none; } code a { color: inherit; }
Open file (34.23 KB 640x480 chb_hibiya033.jpg)
Please Note: I'm going to be continuing with occasionally repopulating our board with old posts. Feel free to post responses to them regardless, as I will generally be watching and can often respond to your posts. But be aware these will all be 'log dumps' as it were from our original board. These bump-bot posts will all have rw@bump.bot as the email sig. Cheers.
continued from >>3400 >FTFY. Don't expect me to shove that boulder up the hill all by myself lad lol. :^) lmao, fair. I do have some Japanese dude's research papers laying around on the subject. So I wanna steal his idea for having a robo-waifu express emotion w/o needing to build it an actual human-like face.
Open file (2.98 MB 300x224 happyfox.gif)
>>3405 Heh. As an engineer and an artist, I find it interesting that this topic of expressing emotions, say, involves (if you'll please forgive my crude and blocky consolidations, there are obviously many, many more considerations within each and every sub-category) -Design : defining (and confirming) the aesthetic and psychological expression of emotions. All within the yuge scope of human verisimilitude. -Hardware : executing on the designs in the physical world, using nothing but the raw tools required to build robots (plastics, metals, servos, gears, pulleys, electronics, etc). All within the yuge scope of human verisimilitude. -Software : executing on the design in the computational world, using nothing but high-level (abstractions, generic/functional programming, etc.) or low-level (signal and line levels inputs/outputs, assembler-tier instruction orderings, etc.), and literally described by mere text. All within the yuge scope of human verisimilitude. Obviously, the reality in the end will come down to many other things, once the 'real world' impinges itself on our robowaifus, not the least of which will be we ourselves and all our human constraints on them. Plus, when you can do things like put the functional equivalent of a dog's tail on the robowaifu's head to help her express emotions properly you just know it's gonna be a wild ride! :-DD Thanks for the document, Fotamon.
>>3406 >codes, all literally described by mere text.*
yeah I gotchu fam something's coming down the pipe at least animal level intelligence I was pissed off for a long time about stupid shit but I'm better now, promise let me know if you're dying for some spoilers
actually let's start with the absolute basics (the question that I was PO'd about) what is the connection between uncertainty, information and life? Entropy! the characterization that defines and informs many such theories. Life maximizes entropy over future states. What does that mean? potentiality. What If you could be homeless tomorrow or a millionaire tomorrow with equal uncertainty? What if you could be homeless or homeless tomorrow? A millionaire or millionaire tomorrow? What does being a millionaire or homeless lead to? This is the crux. Truth told I was dissatisfied. Now I'm not. holler at you Sunday folks
So I'm a little short on time right now so I'll be skipping the theory, anyways let's at least get started: there are two libs you're going need in order to set up your meta-RL maxEnt machine- 1) Erik Laukien's most excellently lightweight and fast sparse coder- https://github.com/ogmacorp/OgmaNeo2 (which btw is fast enough for realtime; TF and PyTorch in general are not) 2) MI toolkit- https://github.com/Craigacp/MIToolbox 3) a quick and dirty test file- mountainCar which should be included in the files (or not- later), I honestly can't remember where I got this from, and uh license reqs and credits will have to come later I guess so go ahead and follow the install instructions, set up your cmake and start a new cpp, for MIToolbox you need to insert a #define COMPILE_C before the includes includes- #include "Entropy.h" #include "ogmaneo/Hierarchy.h" #include "MountainCar.h" ...alright so given that my project files were utterly gutted last week so that I could redesign everything to be more authentically neuro-scientific (complete with short term/long term/working memory) I will actually have to re-code the original project tonight or tomorrow night (protip: don't be me- use git). I'll upload the project files at that point rather than sitting here typing all this shit out sounds good? look forward to seeing if we can get a bot going with this set up- there are also python bindings
>>3931 Hey there Iggy, welcome back. Interesting looking project I'll take a look. It would indeed be nice to have a fast sparse system, so yea. How have things been with you mate?
>>3933 All over the place. There's still too much negativity in my life and I'm sick of it. There's this sort of general malaise and restlessness I need to shake but it's proving... difficult. It hasn't impacted research and development too much overall though so that's good. There have been some major breakthroughs in that regard, just things I've come across or what not. I have a pretty clear idea of where I think I should go with all this. Future's looking up there. This current project is pretty awesome so far. The reinforcement mechanism is, as far as I can tell, ubiquitous and universal to life. It can play Atari at least! I think anyone whose serious about robowaifus is going to be able to appreciate it I think I came in at the wrong time though- it looks like robowaifu is getting its own imageboard or something? lol idk whatever. Anything I missed?
>>3935 You'll be fine lad, just don't quit is the key. >There have been some major breakthroughs in that regard, just things I've come across or what not. >The reinforcement mechanism is, as far as I can tell, ubiquitous and universal to life. Very nice. I hope you set up a repo soon for your work. Neuromorphic computing is probably going to turn out to be the big breakthrough everyone in AI is looking for, so yea. >I think I came in at the wrong time though How could any time be the 'wrong' time heh? :^) >it looks like robowaifu is getting its own imageboard or something? Not to my knowledge. We'll be staying on with Julay as it transitions to a new site (probably sometime this summer). We have a bunker board over on Anoncafe as a backup gathering point if necessary, and there are multiple domains here on Julay. https://julay.world/.static/fallback.txt >Anything I missed? The board gets entirely archived regularly now, so there's that. We lost a little of the board from before, but most of it was saved. Now everything is.
>>3937 Ah okay I see now- and hopefully the new image board software will really help facilitate development. That's really the only thing I'm hoping for at this point, well that and more members ^_^. BTW I don't think I can quit AI at this point, it's been years and even if I stop I always go back, just like I keep coming back to humanoid robotics and I especially can't quit now that something AIXI-esque is coming within reach (aka quantum computing is going to give descriptional complexity the succ). At most I'm thinking of branching out a little into synthetic biology in the future, really interesting to imagine the potential of biology as applied to material science and just the impossible scale and efficiency of biological computation which brings me to: very much agreed neuromorphic computing is going to make a big splash, oh hell yes: like the brain uses 20 watts and the visual cortex alone has a ~650GB/s bandwidth or something crazy like that, and then turn around and look at how even nvidia's top end cards with thousands of cuda cores chew up clock time and spit out utilities bills when training and deploying recurrent and adversarial models and that's why you end up with things like tensorRT. My only wish for neuromorphic computing/AI is that it follows a very flexible and general programming model that allows sparse coding, deep learning, NLP, neuroscientific models etc to all run on the same chip... oh and if it's not some canned, proprietary bullshit that would be nice too (aka it needs to be "done right"- lookin at you IBM smh) About making a repo: I was on the fence about posting anything to github but idk you can make repos private so that it's 'for robos only' or something, that works. All of my code is pretty hacked together, mostly open source but also I don't seem to care where it's from as long as it works, so there's a slight chance of licensing/legal/ethical issues, for instance I don't even know what the license is for the 10 year old nvtt libs In any case the code has been ready since Sunday night, so if you want I can upload a zip with static libs + cmake file, assuming you use already have the requisite openMP and/or I can set up a private repo but that will take longer. Any preferences?
>>4045 >nvidia's top end cards with thousands of cuda cores chew up clock time and spit out utilities bills heh, great way to word it. >repo just use gitlab, i think it's what most of us here are using tbh. i'll leave the lawyering to the lawyers. just license everything you have with MIT license and everything should be good to go. closing the stable door after the horse has bolted and all that you know. sure, you can post your code on catbox.moe or anonfiles or w/e. i already have both the repos you linked d/l'd & building & I'm really big on C++ already, so yea please do. I'm on Arch so I should be able to accommodate it afaict.
>>4045 >AIXI-esque While researching this, I discovered the original author is involved with offering Half a million Euros Prize for Compressing Human Knowledge http://prize.hutter1.net/ The job is to compress a test subset of Wikipedia better than anyone else. Prove it and you can be rich.
>>4046 >https://anonfiles.com/... pressed for time again lmao anyways here's a tiny proof-of-concept which I'll be expanding in the coming weeks >--- I think it'd be best not to include pre-compiled library blobs Iggy. everyone can build their own and install themselves via the repos. Thanks.
Edited last time by Chobitsu on 06/24/2020 (Wed) 08:33:08.
>>4049 Alright, thanks I'll have a look at it. Have a good one, see you again soon I suppose.
See now we're talking right here >>4047 So I thought about it last night and actually the thing /robowaifu/ needs most is to make some paper. That's the flag to rally around. Everything that could fix this board stems from that, and the lack of obvious business opportunities with a good barrier to entry (too high for individuals, low enough for a distributed development team) is thus the sole preoccupation >>4050 so if you installed the libraries the only thing to do is grab cmake lists and main.cpp and mountainCar.cpp/.h there's really not much to look at, I threw it together in like 45 minutes then added some comments and uploaded it here's an interesting facet: you can actually turn the learning off in the step(cs, ...) function and the network will actually still improve on novel tasks; this is more apparent in the ALE anyway I have a lot of work to do catch y'all later
>>4051 Thanks for the details, Iggy. >the network will actually still improve on novel tasks; this is more apparent in the ALE That's interesting. I wonder how that works.
>>4051 >So I thought about it last night and actually the thing /robowaifu/ needs most is to make some paper. That's the flag to rally around. Everything that could fix this board stems from that, and the lack of obvious business opportunities with a good barrier to entry (too high for individuals, low enough for a distributed development team) is thus the sole preoccupation So take up the mantle of 'Business Manager' or something for /robowaifu/ Iggy. This board has always been about herding cats in a general direction. Anons will work on a subset they are currently interested in. For me right now, that means mastering concurrency and parallelism using C++20. For you atm, that apparently means finding business opportunities for the community. "See a need, fill a need" as the old saying goes. Good luck. Whoever can capitalize on this progression to a full-fledged robowaifu market will become a very wealthy man indeed.
brb hell

Report/Delete/Moderation Forms
Delete
Report

Captcha (required for reports and bans by board staff)

no cookies?